Novel technique for high-throughput and power efficient cryptographic primitives

Abstract

Most of internet applications incorporate a security layer in which a hash function is found. These applications are addressing to more and more clients and thus the corresponding server accepts a great number of service requests. In order to indulge these requests the security schemes must have a high-throughput. Furthermore, due to the tendency of the market to minimize devices' size and increase their autonomy to make them portable, power issues have also to be considered as long as the client-side is concerned. In this work the parallelism (or partial unrolling) technique is presented for increasing frequency and throughput of all widely used hash functions - and those that will be used in the future- hash functions such as MD-5, SHA-1, RIPEMD (all versions), SHA-256, SHA-384, and SHA-512 etc. The proposed parallelism technique leads to a 33%- 50% higher throughput comparing to the most competitive implementations and to much lower power dissipation.